Creating an Archives Management System at the University of Maryland Libraries

Encoded Archival Description (EAD) is an XML-based
standard used to encode archival finding aids that reflects
the hierarchical nature of archival collections and that provides
a structure for describing the whole of a collection, as well as
its components (Pearce-Moses 2004).
Archives approach EAD in different ways. The tools available
through the official EAD website at the Library of Congress (
<http://www.loc.gov/ead/> ), such as the EAD
Cookbook , focus on making it possible for archivists to tag
their existing paper/word-processed finding aid content. This
approach works well for smaller institutions, but with over 400
finding aids, manual tagging was not a viable solution at the
University of Maryland.
In order to help define the ideal system for our institution, the
staff of the Archives and Manuscripts Department identified
several main themes. Streamlining workflow was the largest
focus. Finding aids are the natural end result of a series of
archival procedures that begin with appraisal, selection,
arrangement, and finally, description. Since these processes are
related, the creation of a system to help the department
streamline workflow while also simplifying and demystifying
the creation of an EAD document seemed the most practical
course of action. Standardization was another concern. We
wanted to create a system that organized our finding aid data
in such a way that a future generation could easily port it to a
new system. Search capability was a third requirement.
Searching across collections would greatly aid us both with our
reference services and with our processing. Fourthly, we wanted
to create a distinct, usable, and attractive public interface. The department had already created a database in Microsoft
Access to keep track of basic information, such as collection
title , collection size , and location. As a first step, we modified
this database to hold several more collection-related fields and
also added a report that allowed staff to create electronic
accession sheets (the first step in the record-keeping process
for any collection acquired by the department).
Since much of the accession sheet information becomes part
of the later EAD document, the new database tables and
structure were based on the structure of EAD. Many of the
fields were named after their corresponding EAD tags.
The database structure is relatively simple. A main table, named
archdescdid after one of the main components of an EAD
document, contains the bulk of the information. A handful of
smaller tables tie together the <eadheader> information and
the deeper-level descriptive information located in the <dsc>
sections of the EAD document. The biggest challenge was
figuring out a way to design a table that would accurately
represent the "Box Inventory" section of the paper finding aids
so that data entry was simple for staff, but that would also easily
convert into the EAD tag structure.
The decision to use Microsoft Access as the primary database
was based on a number of rationale, although the primary ones
were staff familiarity and widespread availability of the product
within the institution. Several other institutions use web forms
for entering EAD information into a database, and while this
method is very flexible and allows the system to be easily
shared across institutions, it would not allow the department
to carry out some of the other collection management tasks.1
In the absence of a skilled programmer, Microsoft Access
adequately served the purposes of the early phases of the
project. Little to no programming expertise was necessary to
create functional database forms and reports. There were,
however, some weaknesses in the Microsoft Access software
that put the project on hold at a crucial point: the plan to create
the EAD document using the Microsoft Access report features
would not work; the reports could not handle the text in large
memo fields. A model for the conversion from Microsoft Access
to EAD came from the Australian Heritage Document
Management System, which was created by the Australian
Science and Technology Heritage Center . It also used an
advanced Microsoft Access database and ASTHC staff was
helpful in discussing the system. After examining their
approach, the University of Maryland realized that the
assistance of a programmer would be needed to properly extract
the data from Microsoft Access.
The Archives and Manuscripts Department thus approached
Maryland Institute for Technology in the Humanities (MITH)
to assist with the programming support needed, as well as the
project management skills to convert the Microsoft Access
database into a series of outputs (primarily finding aids and
subject guides), as well as create an online publishing system
with a robust search and browse interfaces, and an
administrative management system.
The software itself is comprised of two independent systems:
a converter program written in Java that communicates with
the Microsoft Access database using Java Database
Connectivity (JDBC), and a web application with an XML
Content Management System.2 The web application is based
on Java Servlet API with Model View Controller architecture.
The converter application creates a list of finding aids in the
database and a user can click and generate the EAD-compliant
XML document.
Figure 1: The converter application which transforms the finding aids from
the Microsoft Access database into EAD-compliant XML.
These documents are then uploaded and indexed by the web
application. The web application also generates the subject
guides and finding aids using XSL style sheets. Figure 2. The home page of ArchivesUM, with a pull-down menu listing the
subject guides which are generated through a combination of static HTML
and dynamically-generated content from the database.
Via the administrative interface, the repository editor can
upload, delete, and convert finding aids to HTML. This
pre-processing of the XML document was built into the system
so that the finding aids did not have to be converted to HTML
at the time of the request. Figure 3 shows a result page ranked
in order of relevance. It was decided that in the first instance,
all collections would be represented in the database through an
abstract. As finding aids are converted, they will be made
available through the archive management system. As Figure
3 shows, the interface makes it clear when the finding aid is
available:
Figure 3: A ranked result page indicting which items have finding aids
available.
Generating subject guides proved a greater challenge. Although
it would have been easy to generate the subject guides on the
fly, it was felt that these needed to be converted into static
HTML pages and mounted on the Internet. Subject guides
indexed by Google and other search engines has proved to be
the most popular way for potential users to find the University
of Maryland 's archival resources. Thus, a feature was built into
the administrative interface to create the subject guides through
a combination of static text and abstracts generated from the
EAD document, where <abstract> tags with different "type"
attributes are located.
Various nodes of the <eadheader> and <archdesc> are
indexed with Lucene and a query interface is provided to search
and browse the finding aid.3 The use of Lucene as a search
index enables compound searches for phrases in the box
inventory, collection title, author, scope, and subject fields of
the EAD document.
Figure 4: A search page enabling users to perform complex searches based
on information in different parts of the EAD finding aid.
The Archives and Manuscripts Department staff and MITH
worked together in the development of several XSLT style
sheets for various parts of the website. In many ways, this
proved to be the most difficult task. The hierarchical nature of
the display of a finding aid made design of the final, and most
important, style sheet extremely complicated. Other repositories
provided examples to build from, but since the EAD of the
<dsc> section of a document varies widely from institution
to institution, advanced customization was necessary.
The administrative interface provides an interface to upload an
XSL style sheet, so that the website administrator can change
the design of the finding aids and subject guides. Much of the
software code for this project has been borrowed from
teiPublisher. Moreover, although staff developed the system
for use with finding aids in three of the units within the
University of Maryland's Archives and Manuscripts
Department, staff constructed it with the possibility that other
archival units on campus could use it, as well as staff in
repositories across the University of Maryland system. While
each repository will have its own Microsoft Access database
so it may generate reports unique to its holdings, there will be
one EAD repository, which will give users unprecedented
access to search across archival units and institutions in a way
not possible currently. This paper will thus address the theoretical, practical, and
programming decisions that contributed to the design of this
archival management system.
1. Virginia Heritage Guides to Manuscript and Archival Collections
in Virginia; Online Archive of California.
2. JDBC <http://java.sun.com/products/jdbc/> .
3. Lucene <http://jakarta.apache.org/lucene/doc
s/index.html> .
Bibliography
Dooley, Jackie M., ed. Encoded Archival Description: Context,
Theory, and Case Studies. Chicago: Society of American
Archivists, 1998.
EAD Help Pages - Software Products. EAD Roundtable of the
Society of American Archivists, 2003. Accessed 2003-08-13.
<http://jefferson.village.virginia.edu/ea
d/products.html>
Encoded Archival Description (EAD): Official EAD Version
2002 Website. Library of Congress, 2002. Accessed 2005-03-21.
<http://www.loc.gov/ead/>
Feeney, Kathleen. "Retrieval of Archival Finding Aids Using
World-Wide-Web Search Engines." American Archivist 62.2
(Fall 1999): 206-228.
Heritage Document Management System. Australian Science
and Technology Heritage Center, 2003. Accessed 2003-04-15.
<hhttp://www.austehc.unimelb.edu.au/HDMS/
findingaids.html>
Miller, Fredric, ed. Arranging and Describing Archives and
Manuscripts. Chicago: Society of American Archivists, 1990.
Online Archive of California. University of California, 2004.
Accessed 2005-03-21. <http://www.cdlib.org/ins
ide/projects/oac/toolkit/>
Pearce-Moses, Richard. "Encoded Archival Description." A
Glossary of Archival and Records Terminology. Website:
Society of American Archivists, 2004. <http://www.arc
hivists.org/glossary/>
teiPublisher. Accessed 2005-05-19. <http://teipubli
sher.sourceforge.net/docs/index.php>
Virginia Heritage Guides to Manuscript and Archival
Collections in Virginia. University of Virginia, 2004. Accessed
2004-11-04. <http://www.lib.virginia.edu/vhp
/index.html>